Our principles
We build AI to help people do better work, not to replace their judgment where it matters. These rules apply to every AI feature we design or deliver.
Human in the loop
When an AI output can affect money, health, legal status or a person's rights, a human reviews and approves it. AI suggestions are clearly marked as such.
Data protection
- Client data is used only for the client's project.
- We do not use client data to train public or third-party models, and we choose AI providers and settings that respect this.
- Sensitive data is minimised, access-controlled and, where possible, anonymised.
Quality and testing
Before launch, each AI feature is measured against an evaluation set built with the client. We monitor quality after release and fix regressions.
Transparency
End users are told when they are interacting with an AI system, and can always reach a human.
On this website
Ask AlgorIA on our homepage gives instant, rule-based suggestions in your browser. What you type there is not stored or sent to us unless you choose to send it through the contact form.
Questions
Contact contact@algoria.tn.